Low-latency forward error correction for high-speed serial links

Abstract

Systems and devices can include a port for transmitting data; and a link coupled to the port. The port, in preparation to transmit a data block across the link, to determine a size of a burst of data to be transmitted across the link; determine a plurality of error correcting code words for forward error correction based on the size of the burst of data; interleave each of the plurality of error correcting code words to correspond with consecutive symbols of the burst of data; and transmit the burst of data comprising the interleaved plurality of error correcting code across the link.

What is claimed is: 
     
       1. An apparatus comprising:
 a port for transmitting data; and 
 a link coupled to the port to:
 determine a size of a burst of data to be transmitted across the link; 
 determine a plurality of error correcting codes for forward error correction of the burst of data based on the size of the burst of data; and 
 transmit the burst of data and the plurality of error correcting codes across the link, wherein the error correcting code words are interleaved in the transmission such that three consecutive bytes in the transmission include a first byte comprising a portion of a first error correcting code, a second byte comprising a portion of a second error correcting code, and a third byte comprising a portion of a third error correcting code. 
 
 
     
     
       2. The apparatus of  claim 1 , wherein the port is to:
 determine a plurality of error correcting codes based on the size of the burst of data by determining a number of consecutive symbols the burst of data will utilize; and identify a number of error correcting codes for each of the consecutive symbols of the burst of data. 
 
     
     
       3. The apparatus of  claim 1 , wherein the port is to:
 divide a burst of data into a subset of symbols; 
 determine a number of consecutive symbols forming the subset of symbols that spans the entire number of bits forming the burst of data; and 
 determine a number of error correcting codes based on the number of consecutive symbols. 
 
     
     
       4. The apparatus of  claim 3 , wherein an error correcting code from the plurality of error correcting codes corresponds to one symbol from the number of consecutive symbols. 
     
     
       5. The apparatus of  claim 1 , wherein the port is to:
 determine a bandwidth utilization for the link; 
 determine a length of error correcting code for forward error correction using the bandwidth utilization; 
 encode a physical layer token with the length of error correcting code; and 
 transmit the physical layer token across each lane of the link. 
 
     
     
       6. The apparatus of  claim 5 , wherein the physical layer token is transmitted as a last double word of a previous forward error correction block. 
     
     
       7. A method comprising:
 determining a size of a burst of data to be transmitted across a link; 
 determining a plurality of error correcting codes for forward error correction of the burst of data based on the size of the burst of data; and 
 transmitting the burst of data and the error correcting code across the link, wherein the error correcting code words are interleaved in the transmission such that three consecutive bytes in the transmission include a first byte comprising a portion of a first error correcting code, a second byte comprising a portion of a second error correcting code, and a third byte comprising a portion of a third error correcting code. 
 
     
     
       8. The method of  claim 7 , further comprising:
 determining a plurality of error correcting codes based on the size of the burst of data by determining a number of consecutive symbols the burst of data will utilize; and 
 identifying a number of error correcting codes for each of the consecutive symbols of the burst of data. 
 
     
     
       9. The method of  claim 7 , further comprising:
 dividing a burst of data into a subset of symbols; 
 determining a number of consecutive symbols forming the subset of symbols that spans the entire number of bits forming the burst of data; and 
 determine a number of error correcting codes based on the number of consecutive symbols. 
 
     
     
       10. The method of  claim 9 , wherein an error correcting code from the plurality of error correcting codes corresponds to one symbol from the number of consecutive symbols. 
     
     
       11. The method of  claim 7 , further comprising:
 determining a bandwidth utilization for the link; 
 determining a length of error correcting code for forward error correction using the bandwidth utilization; 
 encoding a physical layer token with the length of error correcting code; and 
 transmitting the physical layer token across each lane of the link. 
 
     
     
       12. The method of  claim 11 , wherein the physical layer token is transmitted as a last double word of a previous forward error correction block.